That's about the only sensible use of their research setup, actually. At least it's not a total waste that way. Because I'm convinced they wasted everyone's time doing that project. Not only there's no new science, but no new engineering either. It's just sad.
And they have essentially added nothing new to what has been already out there for ages. Their only contribution is a tutorial in applying a stochastic modeling package to a not-very-interesting problem. They have limited themselves to on-the-air aspect of the system only, and on top of that all they show is that in TDMA fixed time slot assignment results in better reliability than random slot assignment. In the latter you get collisions and packet loss rate that's 5 orders of magnitude higher.
All this is done in total isolation from real life. They do not even pretend to take into account various forms of interference, reliability of individual components, etc. I'd never let them publish this paper -- it's just regurgitation of well known stuff under pretense of some "fancy" application. It's pathetic, really. Sorry, but I have nothing better to say about it. It's not a useless paper -- I enjoyed reading it, it's a nice if pretentious introduction to stochastic modelling using Prism, but it belongs on their departmental website as a tutorial, and that's about it. Even if they did it using bidirectional communications and force feedback, it'd still be way too far from being a sensible contribution to the field.
If you're happy with reverse DNS that looks like foo-123-54-32.bar.baz.net, and that has nothing to do with the domain the email purportedly comes from... That reverse DNS is of course pretty standard, but it doesn't solve the entire issue. There are mail hosts that will belch if the machine the SMTP connection comes from is not on the list of MXs for the sender's domain or if the MXs for the sender's domain aren't in sender's domain (sad but true). Thus putting foo-123-54-32.bar.baz.net as your MX may solve half the problem only -- there's a reverse DNS, but it's not what it "should" be. And if you cheat, and set say mail1.mydomain.net IN A 11.32.54.123, then of course rDNS doesn't point to your domain and that's it. Yes, there are asses out there that demand that if you send email, all of the infrastructure should be delegated to you and noone else. We decided they weren't an important enough customer and ignored them...
Any ISP will setup rDNS entries if you have a business account
For certain subsets of "any", that is. I'm with two rather large national-scale ISPs, on business class accounts with both of them, and asking for rDNS makes their eyes glaze over and that's the end of that. Our previous ISP, also a large company, was just as inept...
Reverse DNS costs good money: you need an IP delegation, typically. Many ISPs will charge you extra on top of what IANA or wtfever charges for that. It's hard to justify, for a small business, spending $1.5k+ on something that will prevent maybe 5 outbound emails a month from being misclassified...
It's not like Postscript doesn't use dynamic memory allocation and garbage collection and all that. I don't think that it's unthinkable for a printer to speak Smalltalk as a page description language.
There are digital sextants out there you know. Heck, they are even in daily use on satellites. Called star trackers, even. But there are terrestrial versions, too, and they automatically identify the stars in their FOV and all that. Inertially stabilized, even... as if they were designed for ships, kinda!
Reprogramming an ATM card, on the other hand, is pretty much for identity fraud and not much else
Not really. Someone I've heard from is routinely programming virtual credit card numbers issued by his bank's flash app onto plastic, essentially getting single-use cards with spending- and validity limits.
The smart kid can help teach others. My daughter has somehow started doing that on her own during math. At least she'll get to realize not everyone thinks alike, and somehow you have to pitch things differently to different people to get them to understand. That is, I think, an important skill to have. Every bright person should be able to teach, and you can't be good at it without experience. Might as well start early.
I think your friend has massively overpaid on a set of amplified speakers. You can get a Bose set on eBay for under $200 BIN, or about $100 if you're willing to wait a few weeks and snipe your stuff. Even with shipping to Europe it shouldn't be more than USD 250.
He might have (even unwittingly) bought a device that wasn't authorized by Apple. I have yet to see an amplified dock that's licensed by Apple and wouldn't work across their product line -- as long as the connector fits. Perhaps an older dock was only providing 12V charging voltage -- this is an oversight of the designers, then, Apple has nothing to do with it.
One more thing: IIRC, 2D slices of magnetic circuits have analogs in 2D conductive electric sheets. Magnetic field lines are replaced with electric current. So you can visualize free space as some conductive material, and then imagine that Mu-metal on "back of the magnet" being a thin bar placed on such sheet with higher conductivity, and the magnet being two other conductive bars with a voltage source attached to them. You can measure the analog of magnetic field gradient using a voltmeter;)
The analogy of course has limitations, but if you can visualize flow of current in your head, you can visualize magnetic field lines, too.
Mu-metal acts as a short in the magnetic circuit. If it's merely attached to the back of a flat magnet, it won't do much as it's more-or-less an open circuit, magnetically: the magnetic field lines will pass right through it, they have nowhere else to go. If it were to go around the magnet, though, it'd be a whole different thing.
You'll find Mu-metal shields around cathode ray oscilloscope tubes. I think every Tektronix scope I have has one. The metal is fragile in the sense that mechanical stresses from handling can easily kill its performance. It then has to be re-annealed IIRC.
It's all public information, should be easy to find out. Namely, FCC antenna registrations for example. Unless it's a nearby ham, but then it should be easy to see in google earth overhead views or from street view if it covers the area.
My view is skewed because I had some hang of electronics for better part of my life, so if something electrical/electronics breaks down, and it's as simple as a welder, it's a second thought to fix it. With Harbor Freight stuff I usually open it up and retighten all fasteners anyway before throwing the switch for the first time. If there are soldered/crimped connections, I inspect them and repair/replace as needed.
The clamps are used to ground certain structural steel, as required by code in most places for commercial and multifamily buildings. The structural steel becomes grounded, but is not used as a grounding conductor. Key difference of one syllable.
Slashdot Mirror
That's about the only sensible use of their research setup, actually. At least it's not a total waste that way. Because I'm convinced they wasted everyone's time doing that project. Not only there's no new science, but no new engineering either. It's just sad.
And they have essentially added nothing new to what has been already out there for ages. Their only contribution is a tutorial in applying a stochastic modeling package to a not-very-interesting problem. They have limited themselves to on-the-air aspect of the system only, and on top of that all they show is that in TDMA fixed time slot assignment results in better reliability than random slot assignment. In the latter you get collisions and packet loss rate that's 5 orders of magnitude higher.
All this is done in total isolation from real life. They do not even pretend to take into account various forms of interference, reliability of individual components, etc. I'd never let them publish this paper -- it's just regurgitation of well known stuff under pretense of some "fancy" application. It's pathetic, really. Sorry, but I have nothing better to say about it. It's not a useless paper -- I enjoyed reading it, it's a nice if pretentious introduction to stochastic modelling using Prism, but it belongs on their departmental website as a tutorial, and that's about it. Even if they did it using bidirectional communications and force feedback, it'd still be way too far from being a sensible contribution to the field.
Yeah, but it can't be fixed on our end :(
If you're happy with reverse DNS that looks like foo-123-54-32.bar.baz.net, and that has nothing to do with the domain the email purportedly comes from... That reverse DNS is of course pretty standard, but it doesn't solve the entire issue. There are mail hosts that will belch if the machine the SMTP connection comes from is not on the list of MXs for the sender's domain or if the MXs for the sender's domain aren't in sender's domain (sad but true). Thus putting foo-123-54-32.bar.baz.net as your MX may solve half the problem only -- there's a reverse DNS, but it's not what it "should" be. And if you cheat, and set say mail1.mydomain.net IN A 11.32.54.123, then of course rDNS doesn't point to your domain and that's it. Yes, there are asses out there that demand that if you send email, all of the infrastructure should be delegated to you and noone else. We decided they weren't an important enough customer and ignored them...
I think he should provide a link to your post in the reply!
Any ISP will setup rDNS entries if you have a business account
For certain subsets of "any", that is. I'm with two rather large national-scale ISPs, on business class accounts with both of them, and asking for rDNS makes their eyes glaze over and that's the end of that. Our previous ISP, also a large company, was just as inept...
Reverse DNS costs good money: you need an IP delegation, typically. Many ISPs will charge you extra on top of what IANA or wtfever charges for that. It's hard to justify, for a small business, spending $1.5k+ on something that will prevent maybe 5 outbound emails a month from being misclassified...
It's not like Postscript doesn't use dynamic memory allocation and garbage collection and all that. I don't think that it's unthinkable for a printer to speak Smalltalk as a page description language.
Hecklers are usually entertaining when the comedian is good :)
There are digital sextants out there you know. Heck, they are even in daily use on satellites. Called star trackers, even. But there are terrestrial versions, too, and they automatically identify the stars in their FOV and all that. Inertially stabilized, even... as if they were designed for ships, kinda!
That can be done fairly well in software given some mikes and a multichannel sound capture device ;)
You've got me, I'm not sure how far you have to go to "re-anneal".
Reprogramming an ATM card, on the other hand, is pretty much for identity fraud and not much else
Not really. Someone I've heard from is routinely programming virtual credit card numbers issued by his bank's flash app onto plastic, essentially getting single-use cards with spending- and validity limits.
As a shareholder I'd say I don't mind the buzz :)
What has IQ got to do with anything? It's hardly a measure of how good a teacher you are, merely how much effort you put into "passing" the test.
The smart kid can help teach others. My daughter has somehow started doing that on her own during math. At least she'll get to realize not everyone thinks alike, and somehow you have to pitch things differently to different people to get them to understand. That is, I think, an important skill to have. Every bright person should be able to teach, and you can't be good at it without experience. Might as well start early.
I think your friend has massively overpaid on a set of amplified speakers. You can get a Bose set on eBay for under $200 BIN, or about $100 if you're willing to wait a few weeks and snipe your stuff. Even with shipping to Europe it shouldn't be more than USD 250.
He might have (even unwittingly) bought a device that wasn't authorized by Apple. I have yet to see an amplified dock that's licensed by Apple and wouldn't work across their product line -- as long as the connector fits. Perhaps an older dock was only providing 12V charging voltage -- this is an oversight of the designers, then, Apple has nothing to do with it.
I do down to 0402, without any optics other than what comes with my own eyes. I don't know how much longer I'll be able to pull it off, though.
One more thing: IIRC, 2D slices of magnetic circuits have analogs in 2D conductive electric sheets. Magnetic field lines are replaced with electric current. So you can visualize free space as some conductive material, and then imagine that Mu-metal on "back of the magnet" being a thin bar placed on such sheet with higher conductivity, and the magnet being two other conductive bars with a voltage source attached to them. You can measure the analog of magnetic field gradient using a voltmeter ;)
The analogy of course has limitations, but if you can visualize flow of current in your head, you can visualize magnetic field lines, too.
Mu-metal acts as a short in the magnetic circuit. If it's merely attached to the back of a flat magnet, it won't do much as it's more-or-less an open circuit, magnetically: the magnetic field lines will pass right through it, they have nowhere else to go. If it were to go around the magnet, though, it'd be a whole different thing.
You'll find Mu-metal shields around cathode ray oscilloscope tubes. I think every Tektronix scope I have has one. The metal is fragile in the sense that mechanical stresses from handling can easily kill its performance. It then has to be re-annealed IIRC.
It's all public information, should be easy to find out. Namely, FCC antenna registrations for example. Unless it's a nearby ham, but then it should be easy to see in google earth overhead views or from street view if it covers the area.
My view is skewed because I had some hang of electronics for better part of my life, so if something electrical/electronics breaks down, and it's as simple as a welder, it's a second thought to fix it. With Harbor Freight stuff I usually open it up and retighten all fasteners anyway before throwing the switch for the first time. If there are soldered/crimped connections, I inspect them and repair/replace as needed.
Grounding electrode is not a grounding conductor for a branch circuit IIRC.
The clamps are used to ground certain structural steel, as required by code in most places for commercial and multifamily buildings. The structural steel becomes grounded, but is not used as a grounding conductor. Key difference of one syllable.
This link shows typical residential single-family electrical hookup and says nothing about using structural steel as grounding conductor in a circuit.